• DocumentCode
    190828
  • Title

    A stochastic computation based integer DCT implementation in HEVC

  • Author

    Shuaifu Zhang ; Yu Shen ; Canmei Yang

  • Author_Institution
    Sch. of Inf. Sci. & Technol., Univ. of Sci. & Technol. of China (USTC), Hefei, China
  • fYear
    2014
  • fDate
    5-8 Aug. 2014
  • Firstpage
    153
  • Lastpage
    157
  • Abstract
    In this paper, a stochastic computation (SC) based hardware implementation for 4-point DCT in the emerging High Efficiency Video Coding (HEVC) standard was provided. HEVC employs integer DCT with larger transform coefficients than the preceding standards. Hence the multipliers and adders therein are also more hardware consuming. With SC theory applied to DCT hardware design, the circuit implementation is simplified. In DCT implementation the parallel architecture was adopted based on weighted binary generator to improve the whole computing throughput. The impact of different shift lengths on the accuracy of SC multiplier was also analyzed. Based on SC DCT, We propose a modified architecture which greatly improves the whole computing accuracy. The proposed two SC based DCTs are implemented on Xilinx FPGA. They both have more than 50% hardware resource saved and 2 times working frequency compared with traditional methods.
  • Keywords
    discrete cosine transforms; field programmable gate arrays; stochastic processes; video coding; 4-point DCT; HEVC standard; Xilinx FPGA; hardware implementation; high efficiency video coding; integer DCT implementation; stochastic computation; Accuracy; Computer architecture; Discrete cosine transforms; Hardware; Radiation detectors; Standards; DCT; FPGA; HEVC; stochastic computation;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Signal Processing, Communications and Computing (ICSPCC), 2014 IEEE International Conference on
  • Conference_Location
    Guilin
  • Print_ISBN
    978-1-4799-5272-4
  • Type

    conf

  • DOI
    10.1109/ICSPCC.2014.6986172
  • Filename
    6986172